siebrenf/lore

LOng-read Repeat Element pipeline

Long-read Repeat Element pipeline

for PacBio single-cell MAS-seq data

LoRE workflow

Expand minimal workflow

Expand maximal workflow

Underlying MAS-Seq workflow overview

Expand

How install LoRE:

Clone the repository:

git clone https://github.com/siebrenf/lore.git

Create the conda environment:

conda env create -n lore -f lore/requirements.yaml
conda activate lore

Install LoRE in the conda environment:

pip install -e ./lore

How to run LoRE:

Change directory into the LoRE folder.

Activate the conda environment:

conda activate lore

Update the config.yaml.

• Adapters, primers and barcodes for the 5' Kinnex kit can be downloaded by LoRE, or can be placed inside the results directory.
• The results directory (as well as any other directory) can be set in the config. The default results directory is ./results.
• The genome and gene annotation need to be obtained manually. You will need to specify their locations in the config, as well as the symbol for the mitochondria.
  • If the pigeon classify output suggests the number of reads per cell is low, this may suggest the genome and/or gene annotation were insufficiently annotated.
• Additional documentation for most rules (steps) in the workflow can be found in the code.
• Optional outputs (currently) include bigwigs (for track visualization) and a QC report. Both are recommended, but adds (some) computational load.

Test your config:

snakemake --snakefile lore/Snakefile --configfile config.yaml --dry-run

Run your config:

nice snakemake --use-conda --snakefile lore/Snakefile --configfile config.yaml --resources parallel_downloads=1 mem_mb=100_000 -j 60 > log.txt 2>&1

Further reading:

• PacBio docs
• PacBio repos
• PacBio datasets
• PacBio glossary

TODO:

• implement TE/RE detection using the output of either:
  • isoseq_groupdedup (an unaligned FASTA and BAM file)
  • pbmm2_align (an aligned BAM file)
    • current settings:
      • multimapped reads are included: reads are assigned to any number of locations (nice for TEs).
      • unmapped reads are included.
  • repositories of interest to this purpose have been marked below.
• learn more about the pigeon classify filter settings (for the gene level, some filters may be more lenient).
• integrate genomepy to get a genome & gene annotation.
  • figure out the requirements for a "good" reference genome & gene annotation.

Repositories of interest for TE/RE detection:

• de novo Repeat library construction pipelines
  • https://doi.org/10.1186/s12864-021-08117-9
    • https://github.com/kacst-bioinfo-lab/TE_ideintification_pipeline # nice workflow figure
    • could be useful for non-model organisms/strains
• RE/TE pipelines
  • SoloTE
    • easy to use: use pbmm2 output
    • https://doi.org/10.1038/s42003-022-04020-5
    • https://github.com/bvaldebenitom/SoloTE
  • TrEMOLO
    • looks good + snakemake!
    • https://doi.org/10.1186/s13059-023-02911-2
    • https://github.com/DrosophilaGenomeEvolution/TrEMOLO
  • tldr
    • looks easy to use
    • https://doi.org/10.1016/j.molcel.2020.10.024
    • https://github.com/adamewing/tldr
    • works with "partial detection, assembly and annotation of non-reference TE insertions"
  • TELR
    • looks good! pure python
    • https://doi.org/10.1093/nar/gkac794
    • https://github.com/bergmanlab/TELR
    • "relies on the detection of abnormally mapped reads upon the reference genome that can be linked to a TE"
  • teNanoporePipeline
    • may be usable
    • https://doi.org/10.1016/j.isci.2023.108214
    • https://github.com/javiercguard/teNanoporePipeline
    • Tools: RepeatMaster, Dfam
  • LoRTE
    • not usable (python 2!)
    • https://doi.org/10.1186/s13100-017-0088-x
    • dead link